# How we measure, and what we refuse to claim

> The six positions that decide what we will and will not print.

Source: https://citetune.com/methodology

## What we measure, and what we refuse to

We report retrieval, citation and answer presence as separate quantities, because a page that was fetched is not a page that was cited.

We publish no composite AI visibility score. A single number that blends six surfaces with different methods hides the one thing you need — which surface changed, and why.

## Completion rate

Every sample reports prompts attempted and prompts completed for each surface. When a surface fails, it is shown as not measured, never as zero, because zero is a result and a failure is the absence of one.

## Verification versus attribution

A re-check proves a specific defect no longer reproduces on a specific date. It does not prove a traffic lift, and we will not draw that arrow for you.

Any visibility change we observe after a fix is reported next to the verification, with its own dates and completion rates, and labelled as observation rather than consequence.

## Controls

A before-and-after with no control cannot separate your change from the growth of AI search itself. We show intervals and sample sizes so you can see when a difference is too small to act on.